| Age | Commit message (Collapse) | Author |
|---|
|
Here we add and enable by default a datasource for Scaleway cloud.
The datasource quickly exits unless one of three things:
a.) 'Scaleway' found as the system vendor
b.) 'scaleway' found on the kernel command line.
c.) the directory /var/run/scaleway exists (this is currently created
by the scaleway initramfs module).
One interesting bit of this particular datasource is that it requires
the source port of the http request to be < 1024.
|
|
This allows the user to seed NoCloud in a trivial way from qemu/libvirt,
by using a stock image and passing a single command line flag. No custom
command line, no filesystem modification, no bootstrap disk image.
This is particularly handy now that Ec2 backend is discouraged from use
under bug 1660385.
LP: #1691772
|
|
Azure sets a known chassis asset tag to 7783-7084-3265-9085-8269-3286-77.
We can inspect this in both ds-identify and DataSource.get_data to
determine whether we are on Azure.
Added unit tests to cover these changes
and some minor tweaks to Exception error message content to give more
context on malformed or missing ovf-env.xml files.
LP: #1693939
|
|
AliYun cloud platform is now identifying themselves by setting the dmi
product id to the well known value "Alibaba Cloud ECS". The changes here
identify that properly in tools/ds-identify and in the DataSourceAliYun.
Since the 'get_data' for AliYun now identifies itself correctly, we can
enable AliYun by default.
LP: #1638931
|
|
If the user configured:
datasource_list: ["Ec2", "None"]
then ds-identify would write
datasource_list: ["Ec2", "None", "None"]
which would break the logic to avoid warning.
|
|
This adds several unit tests for ds-identify, and fixes a bug
in Ec2 detection that I found while writing these tests.
The method of testing is to use the ds-identify code as a shell
library. The TestDsIdentify:call basically does:
* populate a (temp) directory with files that represent what
ds-identify would see in /sys or other locations it reads.
* create a file '_shwrap' that replaces the 3 programs that are executed
in ds-identify code path. It supports setting their stdout, stderr,
and exit code.
* set the default policies explicitly (DI_DEFAULT_POLICY) so we can
support testing different builtins. This is necessary because the
Ubuntu branches patch the builtin value. If we did not explicilty set
it, then testing there would fail.
* execute sh to source the script and call its main.
|
|
Apparently signals were crossed when this implementation was done.
Cloud-init was reading 'platform' in the environment of pid 1, but
nova-lxd was setting 'product_name'.
The fix is being made here in cloud-init to instead read product_name.
LP: #1685810
|
|
We were checking /var/lib/cloud/openstack/latest/meta_data.json instead
of /var/lib/cloud/seed/config_drive/openstack/latest/meta_data.json.
LP: #1673637
|
|
The reading of MAAS datasource configuration was simply broken.
it was looking in /etc/cloud/*maas*.cfg rather than
/etc/cloud/cloud.cfg.d/*maas*.cfg.
along side here there is also:
* doc improvement on check_config
* remove the path restrictions when searching for values in both
maas and ovf_vmware_guest_customization. that was done to improve
performance as check_config's parsing is slow.
* change to maas to search all config files rather than restricting
to a subset as it tried before. that was done for
* better variable names.
- rename path_cloud_confd to path_etc_cloud
- PATH_ETC_CLOUD: /etc/cloud
- PATH_ETC_CI_CFG: /etc/cloud/cloud.cfg
- PATH_ETC_CI_CFG_D: /etc/cloud/cloud.cfg.d
LP: #1677710
|
|
If the only the None datasource was listed in datasource_list, then
ds-identify would write a cloud.cfg witih:
datasource_list: [None, None]
The fix is to just append None if the list only has None.
|
|
OpenStack clouds installed with RedHat RDO have the nova product
configured in /etc/nova/release to be 'OpenStack Compute' rather than
upstream nova default of 'OpenStack Nova'.
This was first reported on Finnish provider Nebula (http://nebula.fi).
LP: #1675349
|
|
While documentation indicates that the smbios product name should
contain 'Google Compute Engine', experimentation and bug reports
indicate that is not always the case. The change here is to change
the check for GCE to also consider a serial number that starts with
'GoogleCompute-'.
Also, ds-identify was not currently searching for GCE if no config of
datasource_list was found. Most images have a datasource_list defined.
So update the list to include GCE.
LP: #1674861
|
|
This is thie cloud-init part of a fix to allow nova-lxd to provide
config drive data. The other part will be done in nova-lxd.
The agreement here is that nova-lxd will copy the contents of the
config drive to /config-drive in the container.
LP: #1673411
|
|
The path for checking presence of Bigstep datasource was simply wrong.
Set the correct path.
LP: #1674766
|
|
The script is written to have the protection of disabling filename
expansion (set -f) and explicitly enabling expansion when needed.
However, the check_config function failed to disable it after enabling.
|
|
Change policy so that 'report' can be overridden.
In xenial we had set the builtin default to be 'report', expecting
that Ubuntu core would install config that changed it to 'search'.
However, if report was already set, there was no way to unset it.
The change here is to make 'report' basically 'search-dryrun', so
that one or the other can be set.
The other change here is that report would actually exit disabled
if it did not find a datasource and notfound=disabled. That was
unexpected and would turn cloud-init off, which is not what we wanted.
Additionally, consistently use 'enabled' or 'disabled' versus
'enable' and 'disable'.
LP: #1669949
|
|
Rather than having the dscheck_Ec2 just know the setting, move
it up to a more formal declaration. This will make it look more
clean when a distro carries a patch to change it to warn.
|
|
On a 'not found' result, was not being written at all.
That had the unintended effect of '--force' not working.
Now, on a 'not found' result:
- if reporting: write the list as found (with just 'None').
- if not reporting: only report that there was nothing found.
this means that the warning cloud-init will write about ds-identify
failing to find a datasource will be written, but cloud-init will
still search its fully configured list.
|
|
Now, when ds-identify runs in report mode, it still writes to
/run/cloud-init.cfg as search does, but it will namespace the
result under the top level 'di_report' entry.
|
|
In the interest of speed I had skipped the parsing of
/etc/cloud/cloud.cfg for the ec2 strict_id setting. In hindsight
it seems reasonable for people to put settings there.
|
|
ovf_vmware_guest_customization defaults to true in cloud-init, meaning
that such customization is disabled. We just missed a return value
causing ovf_vmware_guest_customization to effectively default to on.
Also, when looking for setting look at /etc/cloud/cloud.cfg.
This had been omitted in interest of performance, but we should
be looking there.
|
|
cloud-init by default sets 'disable_vmware_customization' to True.
So in ds-identify, we will enable the ovf datasource if:
- virt is vmware
- 'libdeployPkgPlugin.so' exists as installed by vmware-tools or
open-vm-tools.
- disable_vmware_customization is configured to True
|
|
Brightbox will identify their platform to the guest by setting the
product serial to a string that ends with 'brightbox.com'.
LP: #1661693
|
|
Based on the setting Datasource/Ec2/strict_id, the datasource
will now warn once per instance.
|
|
ds-identify will now read this setting, and thus allow the user
to modify ds-identifies behavior via either:
1. builtin setting here cloud-init/ds-identify builtin
2. ds-identify config (/etc/cloud/ds-identify.cfg)
3. system config (/etc/cloud/cloud.cfg.d/*Ec2*.cfg)
4. kernel command line (ci.datasource.ec2.strict_id=true)
|
|
A check function that returns found or maybe can also now
return config that will be written to the resultant /run/cloud.cfg.
They do so by setting the variable _RET_excfg.
|
|
This just adds checking of the Ec2 seed directory.
|
|
The following can have cause issue:
FOO="bar ; wark"
showit() {
local b=$FOO
echo $b
}
4: local: ;: bad variable name
The answer is just to use more quotes.
|
|
Just remove some examples that are no longer valid.
|
|
This makes ds-identify run only once. Previously it would run multiple
times each boot as the generator would run more than once. This is
potentially dangerous, in that running again might find more attached
disks. However that is really only a "lucky" fix if it happens to
result differently than the first run.
Additionally, we now log the uptime that we started and ended at.
|
|
has_fs_with_label regressed when refactoring to not have leading
and trailing , in DI_FS_LABELS.
LP: #1663735
|
|
The nocloud datasource specifically would look for ds=nocloud or
ds=nocloud-net (often augmented with 'seedfrom') on the kernel command line.
Fix to return DS_FOUND in that case.
LP: #1663723
|
|
Nova lxd will now put the environment variable 'platform' into
pid 1's environment to the value 'OpenStack Nova', which is the same as
you would find in kvm guests.
LP: #1661797
|
|
aarch64 does support dmi, but OpenStack does not populate guests
with this information, and there are currently bugs in qemu preventing
it from working correctly see bug #1663304 for more information.
So, for the time being, pretend as if there is no dmi data on aarch64,
which will make it enable cloud-init even when no datasources are found.
|
|
ds-identify is run here from the generator. If ds-identify does
not see any datasources, it can completely disable cloud-init.
The big value in this is that if there is no datasource, no python
will ever be loaded, and cloud-init will be disabled.o
The default policy being added here is:
search,found=all,maybe=all,notfound=disabled
That means:
- enable (in 'datasource_list') all sources that are found.
- if none are found, enable all 'maybe'.
- if no maybe are found, then disable cloud-init.
On platforms without DMI (everything except for aarch64 and x86),
the default 'notfound' setting is 'enabled'. This is because many of
the detection mechanisms rely on dmi data, which is present only on
x86 and aarch64.
|
